The function samples points or patches on a categorical raster layer.
sample_raster(
raster,
class,
nb_pts,
dist_min = 0,
edge_size = 0,
by_patch = TRUE,
neighborhood = 8,
surf_min = 0,
prop_area = TRUE,
step_max = 1000,
output = "df",
desc = TRUE
)A list of object(s) with one or several elements according to the
output and desc arguments.
A RasterLayer object corresponding to a categorical raster layer
An integer value or vector with the value(s) corresponding to the code values of the raster layer within which points will be sampled.
An integer value indicating the number of points to be sampled
An integer value indicating the minimum distance separating the sampled points (default = 0).
An integer value indicating the width of the edge of the raster layer which is ignored during the sampling (default = 0). It prevents from sampling in the margins of the study area.
A logical value indicating whether contiguous patches with cells having the same code value are delineated prior to sampling (default = TRUE). It prevents from sampling several points in the same contiguous patch.
An integer value indicating which cells are considered
adjacent when contiguous patches are delineated (it should be 8
(default, Queen's case) or 4 (Rook's case)). This parameter is ignored
when by_patch = FALSE.
An integer value indicating the minimum surface of a patch
considered for the sampling in number of raster cells. This parameter is used
whatever the by_patch argument is. Default is 0.
A logical value indicating whether sampling in large patches
is more likely (default = TRUE). If by_patch = FALSE, this parameter
is ignored. When prop_area = TRUE, the probability to sample a given
patch is proportional to its area.
An integer value indicating how many sampling steps are performed to identify a point set satisfying all the conditions before returning an error.
A character string indicating the type of returned output:
'data.frame': A data.frame with three/four columns:
ID: The point or patch centroid ID
x: The point or patch centroid longitude
y: The point or patch centroid latitude
area: The area of the sampled patch (only if by_patch = TRUE)
'pts_layer': A SpatialPointsDataFrame layer corresponding
to the sampled point (points or patch centroids)
'poly_layer': A SpatialPolygonsDataFrame layer corresponding
to the sampled patch polygons
A logical value indicating whether the result should be
described or not (default = FALSE). If desc = TRUE, then the Gini
coefficient of the distances between points and of the patch areas (if
by_patch = TRUE) is computed with the gini_coeff.
An histogram of the link weights is also described.
P. Savary